I believe some of this is the extremely small enclosure volume.
There are appears to be some sort of DSP-related group delay going on that has effects that are beyond simple equalization. Although the enclosures are relatively small, the driver–enclosure combination can still be designed to work reasonably well, but of course sensitivity will be sacrificed. That's all covered by Thiele, Small, Keele, Bullock, amongst others, in the many papers on closed box and vented box systems.
I am not a simulation expert but it looks like you are using a 0.707 QTC sealed enclosure for your baseline which the SVS subs are not even close to.
I think that there may be some misunderstanding of what was being portrayed. In the example that was presented, the choice of Qtc = 0.7071 as the target 2nd-order high-pass response was purely illustrative. It was chosen to show that if the electrodynamic system is minimum phase then applying large amounts of EQ can change the system response to match a different one: in magnitude, phase, and group delay response. This type of behaviour underpins the use of the Linkwitz Transform with closed boxes. The Qtc = 0.7071 system was set as the target, as its response is well known and well understood. This example showed that, even when high levels of EQ are applied, the group delay simply matches that of whatever target is chosen. High levels of EQ
do not necessarily lead to higher levels of group delay. The point of the exercise was merely to demonstrate that behaviour with a concrete example.
The example can be modified to suit the target response as found on an SVS sealed subwoofer. The difficulty of that approach is that we don't have data readily at hand on what the theoretical characteristics are. I suppose VituixCAD or XSim could be used to do an inverse transformation to compute the minimum-phase phase response. However, that is an additional step that somewhat further complicates matters.
The original analysis showed that high-levels of EQ do not necessarily lead to excess levels of group delay beyond that displayed by the natural response of the equivalent non-equalised system. If additional amounts of group delay are present, then it would be interesting to know and understand their cause. Presently, apart from such general labels as "the signal processing chain", I am left in the dark.
The interaction with the air inside these small sealed enclosure, both compression and shearing, can be a large source of non-linearities.
There is certainly going to be quite a bit of compression of the air taking place. Have you studied the stiffness versus compression curve of just such an enclosure? You may be surprised that the air can be relatively linear in its response, especially when compared with the intrinsic nonlinearities of the woofer's suspension. Are you able to elaborate a little on how "shearing" is a large source of non-linearities? I am curious about the mechanisms involved.
These small enclosure aggressively DSP'd subs are designed primarily to fill in below small satellite mains speakers with moderate SPL and for that role they can work quite well and the group delay is likely to be a very minor issue.
Looking at the less than stellar step responses of some of these subwoofers, they may not be as good for the task at hand than would normally be expected. But, being adequate at producing moderately high SPLs is probably sufficient for the intended use case. Of course, some listeners will prefer a better overall design, and that's where the larger-volumed DIY sealed subwoofers come into play on a regular basis.
However, when trying to integrate them with multiple different subs and mains the group delay can really complicate integration.
What you have found is definitely a problem area that one needs to be aware of. You're own measurements, and those of others, have demonstrated that some subwoofers have unusually high levels of group delay. And, as you have found, mismatches between subwoofer group delays can play havoc with the summed response, which is a point that needs to be kept in mind when blending different subwoofers together in a mult-subwoofer installation.
Controversially / subjectively in my experience if you are trying to chase high SPL "slam" from a kick drum or similar I have had better luck with subs that show better time domain behavior for which group delay is a good indicator / proxy.
The transient responses of some of these subwoofers are hardly clean in the textbook sense that their magnitude response might lead us to believe that they would be capable of. There is an awful lot of ringing going on, which would hardly be conducive to accurately reproducing what the original sound pressure waveform looked like. Providing details of a subwoofer's transient response, above and beyond the simple magnitude response as a function of frequency, is something that most manufacturers seem to avoid.
Thanks to some detailed measurements provided by independent testers, we are aware of the potential deficiencies. If one is able to measure the group delay of a subwoofer system, then it is also entirely possible to directly look at things such as the step response and the pulse response by taking some appropriate measurements. Why, even some low-frequency 25-Hz square waves would be fun to see!
